Global Meta-Analysis Integrated with Machine Learning Assesses Context-Dependent Microplastic Effects on Soil...

Area of Science:

• Environmental Science
• Soil Science
• Ecotoxicology

Background:

• Microplastics (MPs) in soil can alter microbial biomass and nutrient cycling.
• Understanding these effects is crucial for soil health and ecosystem function.

Purpose of the Study:

• To quantify the global effects of MPs on soil microbial biomass carbon (MBC) and nitrogen (MBN).
• To identify key factors influencing MP impacts using meta-analysis and machine learning.

Main Methods:

• Global meta-analysis of 90 studies (710 MBC, 354 MBN observations).
• Integration of machine learning (XGBoost) to model nonlinear responses and threshold effects.
• Comparison of field versus controlled experimental conditions and different plastic types.

Main Results:

• Controlled experiments showed significant increases in MBC (9.6%) and MBN (10.4%), unlike field studies.
• Biodegradable plastics (PBAT, PLA) had stronger effects than conventional polymers (PE, PP, PS, PVC).
• Temperature significantly influenced MP effects, potentially decoupling carbon and nitrogen cycles; optimal effects observed for medium-sized MPs at high concentrations in intermediate fertility soils.

Conclusions:

• MPs can stimulate soil microbial biomass, but effects are highly context-dependent.
• Warming temperatures may alter MP impacts on soil carbon and nitrogen cycling.
• Machine learning revealed nonlinear responses and thresholds, highlighting complex interactions.